Project Summary The goal of this grant is to develop enabling technology and systems that address fundamental limitations in microsurgery with a specific focus on vitreoretinal surgery. Due to the inherent micro-scale and the fragility of the neurosensory retina, vitreoretinal surgeons can be challenged by physiological hand tremor where the tremor amplitude is larger than retinal structures, delicate movements that are below tactile sensation, and multiple cognitive decisions that are required when executing high-risk movements, such as during retinal vein cannulation (RVC). Nevertheless currently vitreoretinal surgery is at the limits of human physiological performance and lacks the adequate technology that could further improve the technical performance. This situation is less than optimal and can significantly benefit from the recent advances in medical robotics, sensor feedback and human machine interface design. Robotic assistance may be ideally suited to address common problems encountered in the performance of the demanding micromanipulations in retinal microsurgery. We propose a robotic system with enhanced real-time multisensory feedback that assesses multiple points of instrument contact located both inside and outside of the eye. Our comprehensive system will enable the surgeon to manipulate tools based on quantitative feedback that will prevent mechanical injury by implementing safeguards against the application of excessive and previously unmeasurable forces at the eyewall and the tool tip. Our aims are: (1) Develop and demonstrate in vivo position/force hybrid control algorithms for enabling real- time high-fidelity sensorimotor capabilities at the sclerotomy for safe robot-assisted vitreoretinal microsurgery: real-time sensorimotor capabilities at the sclerotomy will be uniquely used to control the robot through a machine learning method that adaptively learns a nonlinear mapping from user behavior to sclera-force/position and predicts unsafe motions; (2) Develop and demonstrate in vivo force-input control algorithms for enabling real- time high-fidelity sensorimotor capabilities at the tool-tip for safe robot-assisted vein cannulation: real-time tool- tip-to-tissue interaction force sensing and non-linear robot control algorithms based on observing the user behavior will be used to control the tool-tip position and force and to prevent entry into subretinal areas during RVC; (3) Demonstrate safe robot-assisted RVC in rabbit model in vivo: real-time, position/force hybrid control algorithms based on dual-point (tool-shaft and tip) information fusion will provide sensorimotor guidance of surgical maneuvers during RVC. Statistically significant results in vivo, in clinically realistic conditions will demonstrate the feasibility of our approach. This highly innovative system will enable surgeons to perform maneuvers in a tremor free environment with a higher level of precision than previously possible and with the ability to sense forces on a scale that have been previously imperceptible. We envision this development as a logical next step in the integration of man, machine and computer for the performance of unprecedented microsurgical maneuvers.